Statistical Analysis Stack Exchange
Statistical Analysis Stack Exchange

Questions tagged [huber]

13 questions
5
votes
2 answers

How to choose delta parameter in Huber Loss function?

In Huber loss function, there is a hyperparameter (delta) to switch two error function. Currently, I am setting that value manually. But, I cannot decide which values are the best. So, how to choose best parameter for Huber loss function using my…
Dennis Thor
  • 65
  • 1
  • 6
4
votes
1 answer

Tradeoffs of robust mean measures (trimmed, Huber, cosh, etc)

After recently having delved into the world of robust measures (for location, mean being the classical case), I have had difficulty understanding robust measures' core dynamic. Basically, what are you giving away in order to become more resilient…
Coolio2654
  • 642
  • 5
  • 17
3
votes
0 answers

Why is there no improvement when training Xgboost with pseudo-Huber loss?

In this StackOverflow post I asked if there was something wrong with my syntax when training an XGboost model (in R) with the native pseudo-Huber loss reg:pseudohubererror, since nor training or test error improve (remain constant). There doesn't…
PaulG
  • 793
  • 2
  • 10
2
votes
0 answers

In general, how to determine the weight function of Robust regression

I think the question is clear from the title. How the weight function for example in Huber is calculated? Is it by differentiating the objective function?
jeza
  • 1,527
  • 2
  • 16
  • 37
2
votes
1 answer

Huber Loss on top of Cross Entropy

I know that the Huber loss is usually applied on top of the L2 loss in order to prevent exploding gradients. Does it make sense to use the Huber loss on top of the cross entropy loss, though? I have a feeling that it is not very sensible, but a more…
niko
  • 288
  • 2
  • 8
1
vote
0 answers

Reporting robust means and SD to make more comparable to robustbase::glmrob results

I need to report the mean, and SD of two groups for a specific variable, as well as the odds ratio of the two groups using robust logistic regression. The variable does have outliers. I instead want to report the robust mean and sd with a method…
am313
  • 51
  • 4
1
vote
1 answer

Question about Huber loss when k=0 in Casella and Berger

In Casella and Berger (page484), the following Huber loss is defined. Then on the next page, Table 10.2.1 shows the Huber estimator for different $k$: In particular, $k=0$ gives the median, which doesn't make any sense. As when $k=0$, the Huber…
T34driver
  • 1,608
  • 5
  • 11
1
vote
0 answers

Minimizing Huber Loss

Huber Loss is given as follows: I'd like to proof $\gamma=median\{y_1,...y_N \}$ minimizes the Huber Loss so i've taken its first derivate for $\gamma\neq y_i$: I've tried to proof that first derivate change its signal before and after the median,…
Davi Américo
  • 737
  • 1
  • 11
1
vote
0 answers

Existence of Huber M-estimators

I am working on a paper about optimization using the Huber's Loss function, which is defined as: \begin{equation} \psi(x)=\begin{cases} \frac{x^2}{2\gamma},& \text{if } \lvert x\rvert\leq\gamma\\ \lvert x\rvert-\frac{\gamma^2}{2},&\text{if } \lvert…
Deniz Akkaya
  • 11
  • 1
1
vote
1 answer

Scikit-learn: How to normalize Huber regressors?

In scikit-learn the Ridge regression estimator has a normalize parameter that normalizes the regressors. I found that it was necessary to set this to True to get a reasonable fit to my data when using higher degree polynomial features (it provided…
mrossman
  • 113
  • 3
1
vote
1 answer

functional differences between using huber loss and winsorizing/trimming

Curious what the functional differences are between using a Huber loss function/ regression and Winsorizing data and then running a classic least squares regression. Will the resulting outputs be roughly the same? Assuming no gross outliers? Are…
Michael
  • 2,180
  • 4
  • 23
  • 32
0
votes
1 answer

Convex set of huber's contamination model

In the celebrated Huber's robust estimation paper, he considered the following model $x_i \sim (1-\epsilon) P_\theta + \epsilon G$ where $P_\theta$ is assume to be standard normal. Under this model, data is contaminated by some unknown distribution…
cccfran
  • 55
  • 5
0
votes
0 answers

How is the parameter „huber_alpha“ defined in H2O Deep Learning?

In the documentation of the H2O software (http://docs.h2o.ai/h2o/latest-stable/h2o-docs/data-science/deep-learning.html) it says for the parameter “huber_alpha: Specify the desired quantile for Huber/M-regression (the threshold between quadratic and…
user4556577
  • 23
  • 2